Because Uber hired experienced people from CMU (who probably also brought gigabytes of files with them).

Uber hired an entire experienced department from CMU.

Uber acquired Otto around 08/2016. I don't think this explains it.

The lawsuit from Waymo specifically calls out good LiDAR as a component that Uber had not successfully developed prior to acquiring Otto.

Who knows maybe they had somebody else from the Google self driving car project to steal self-driving car secrets earlier. Based on what this Levandowski guy did the industrial espionage may go unnoticed. I'm wondering if Waymo will require Uber to reveal schematics of their self-driving car project as part of the law suite.

I have doubts what was stolen was actually any of the secret sauce. An interface board for a lidar unit is probably one of the most simple things on the list.

The actual self driving software, and more importantly, all of the collected data from the waymo fleet would have been the key.

Not so much an interface board as a whole new tested design for a LIDAR unit, including a unique patented optics setup and laser driver circuit, according to the complaint. Also testing, manufacturing, and characterization procedures and results and information on suppliers for the parts required. The PCB was just the component whose accidental disclosure lead them to conclude that Uber and Otto were using the stolen design. Since the PCB apparently dictates the position and orientation of the laser diodes and sensors, presumably it would only be useful if they copied the whole thing.

Interesting. Is the specific Lidar unit really that big a differentiator? I understand they aren't cheap or simple, but it seems odd that each self driving car company would want to design their own. I would guess you would rather have some healthy ecosystem of suppliers...Velodyne, etc.

> it seems odd that each self driving car company would want to design their own

They don't want to do this, so, if they do, it's because they had to do this.

Practically all of the current sensor suites are expensive, bulky, and power hungry. If you want them on lots of cars, you need to reduce all 3 of those characteristics dramatically.

Frankly, I wish this "patented" (does that word mean anything at all these days ?) tech gets into China, and gives all us hackers cheap Lidars to play with.

Uber launched their beta in Pittsburgh just before Otto acquisition. This doesn't explain it. It explains how they quickly wanted to build their own lidar.

Oh come on. CMU had its own DARPA contenders, and where on earth is this grand institution you ask ? "Pittsburgh" !

Yes. What does this have to do with Levandowski?